This invention relates to signal processors. More particularly, this invention relates to a signal processor having means for converting an analog signal into a digital signal with minimum bias distortion.

A simple optical disc recorder/player may employ a first laser light to write data on an optical disc by burning holes or pits in the medium of the spinning disc. A second laser light recorder/player is usually employed to read the recorded data by reflecting light from the disc in places where a hole or pit has been burned away during recording. The reflected light passing through a set of optics is then applied to a photodetector array where the signals are detected and processed for subsequent utilization.

The playback signal from the optical disc is derived from a laser which has a finite beam size and is proportional to the convolution of the beam and the pits along the tracks of the optical disc. As the minimum pit size in the disc approaches the diameter of the laser beam, the pulse width of the recovered signal approaches the rise time, and the recovered signal approaches the shape of a sinusoidal wave. This signal is mixed with a very low frequency aperiodic thermal drift signal and with a periodic drift-like signal caused by changes in the reflectivity of the disc material.

Theoretically, by dc-coupling and linearly amplifying the signal from the photodiode preamplifier which are elements of the photodetector array, the location of the 50% point of the recovered signal is preserved but the signal is contaminated by noise for any channel modulation code with or without dc content. The use of ac coupling causes the location of the transition of the 50% point to become distorted for a channel coded signal with dc content. The amount of distortion is directly proportional to the amount of dc content of the code.

The previously used method for converting an analog signal to a digital signal was to pass the analog signal through a low pass filter to establish an average dc level. This dc level was then applied to a comparator used as a zero crossing detector or slicer to provide a digital output signal. This method leads to inaccuracies in the disc reading apparatus because the dc content of the input signal introduces an offset in the output signal from the low pass filter resulting in timing errors in the output of the slicer. These timing errors result from the fact that the analog signal from the optical disc is mixed both with very low frequency aperiodic thermal drift and with a periodic drift-like signal caused by changes in the reflectivity of the optical disc material.

From DE-A-2208310 it is known to apply the average peak-to-peak voltage of the analog signal to a comparator used as slicer to provide a digital output signal. In this known method the slicing level varies with the said very low frequency aperiodic thermal drift.

To correct for the timing errors and thermal drift of prior art signal processors, the present invention as claimed provides means to detect the average peak-to-peak analog signal and to pass the average of this peak-to-peak signal through a low pass filter with a cutoff above the periodic drift-like signal. The output of this filter is used as a reference voltage to a comparator which is used as a slicer and which provides a compensated digital signal output. At the same time, the average peak-to-peak signal is fed to a differential amplifier and another low pass filter having a cutoff frequency below the frequency of the periodic drift-like signal. The resulting signal is negatively fed back to the input to be summed and amplified with the input analog signal.

In the drawing:

• Figs. 1A, 1B, and 1C show the relationship between the pit size in the optical disc and the recovered signal under ideal conditions where the pit size is greater than the laser beam diameter;
• Fig. 2 shows the waveforms of an ideal signal detector in which the rise and fall times of the output signal have steep slopes;
• Fig. 3 shows waveforms of a Miller code signal as effected by the high pass filter;
• Fig. 4A shows waveforms of ac coupling versus dc coupling, Fig. 4B is an expanded view of the area A of Fig. 4A;
• Fig. 5 shows a schematic of the static threshold detector;
• Fig. 6 shows the recovered signal amplitude fluctuation caused by variation in the disc reflectivity;
• Fig. 7 is a block diagram of the dc coupled signal processor of the invention;
• Fig. 8 shows the waveforms, the input analog signals, the compensated analog signal and the output digital signal of the dc coupled signal processor;
• Fig. 9 is a block diagram of a signal processor similar to Fig. 8 with the addition of a gap detector, a timing circuit and a selection circuit for playing back a pregrooved optical disc; and
• Fig. 10 shows the waveforms generated by the timing circuit and the sector gap of the signal processor of Fig. 9.

Referring to Figs. 1A, 1 B and 1C, the size of the pit burned into the optical disc during recording is shown in Fig. 1A. As shown in Fig. 1A, the digital information is encoded at the halfway transition point of the recorded signal. Fig. 1 B shows the diameter of the laser beam. Fig. 1C shows that as the laser beam approaches the minimum pit size the pulse width of the recovered signal approaches the rise time causing the recovered signal to approach the shape of a sinusoidal wave. This is undesirable because the output signal from the signal processor becomes distorted.

Fig. 2 shows an ideal signal detector circuit which generates a TTL compatible digital signal from recovered signals. Of course, the ideal detector cannot be achieved and it is the purpose of the present invention to approach a signal output which closely resembles the ideal signal detector.

An analysis of the amount of bias distortion in a signal processor for a Miller modulated signal is discussed below with the description of a static dc coupled threshold detector and a dynamic dc coupled threshold detector. Bias distortion in the environment of a signal processor may be defined as the amount of lengthening or shortening of the output digital pulse caused by deviations of the threshold level away from the half amplitude level of the analog input signal.

In Fig. 3, waveform A shows a signal at the Miller Modulator output with maximum dc content. Waveform B is the dc coupled normalized playback signal after amplification and equalization. The dc content of the signal is shown by dashed lines at ±1/3 amplitude of the signal. The content of an ac signal is defined as the average voltage of the signal. The effect of the high pass filter on waveform B is shown in waveform C where _T, is the time constant of the high pass filter, having low frequency cutoff of f_n=(2πτ)-L For the purpose of analysis and for deriving a formula for calculating edge distortion of the Miller signal, one cycle of the signal is shown for ac and dc coupling in Fig. 4(a). An expanded view of the area A is shown in Fig. 4(b). By analyzing the right-angle triangle formed by line ac with the y and t coordinates, Δt, the amount of distortion at the transition can be calculated.

The slope M₁ of line DC can be approximated bywhere A is the peak amplitude of the dc coupled playback signal, and t_r is its rise time. The slope (M₂) of the line AC can be calculatedwhere M₃ is the slope of dc content at the output of the high pass filter,Since M₁ is typically much larger than M₃, (e.g. for a 4 rps optical recorder, M₁=100 M₃), and M₂≃M₁. The peak transition distortion, Δt can then be calculatedFor the Miller Signal, the pulse duration can be 1T, 1.5T or 2, thereforewhere a=1, 1.5 or 2.

As an example, for a 4 rps optical recorder (a data rate of 1.94 Mbits) using a high pass filter of 15 kHz, with a channel bandwidth (fc) of 1 MHz, then, T=515 nsec. yielding Δt=10.8 nsec. For Miller code, the maximum transition error margin is T/4, i.e. 129 nsec, hence the 15 kHz high pass filter removes 8.4% of this margin.

A static dc coupled threshold detector consists of a voltage comparator and a threshold control as shown in Fig. 5. The threshold voltage V_th is adjusted to the midpoint of the single ended signal V, for the detector to perform properly. Using dc amplification from the photodiodes to V,, the amount of thermal drift (Avd) of a differential pair of bipolar transistors can be calculatedper degree centigrade temperature change of the photodiode, where A=voltage amplification from photodiodes to V,. As an example, let the useful signal from the photodiodes by I m Vp-p, with A=700, then V_I=700 mVp-p, and Δvd≅2.2 mV/°C.

For a 20°C temperature change (ΔT), the expected drift of V, is 44 mV. Since the disc channel bandwidth is limited and the signal has a finite rise time of t_r, the amount of timing error (Δt) iswhere ΔV_t=change in threshold voltage level=ΔvdxΔt.

For a 4 rps optical recorder, t,=0.35 usec. Then for ΔT=20°C, the timing error due to thermal drift is

The static detector requires a readjustment of the threshold level to account for different signal amplitudes due to disc to disc and player to player variation.

Further, even for the same disc, the reflectivity varies from one angular position to another, causing amplitude variation at the disc rotational rate. As shown in Fig. 6, timing error occurs everywhere except where the threshold is proper (A, B, C, D).

The maximum timing error iswhere ΔV =peak signal variation of V, as shown in Fig. 6. In summary, static dc coupled threshold detection is not useful because the timing errors re not with permissible limits.

The dynamic signal dc coupled threshold detector of the invention shown in Fig. 7 automatically compensates for the thermal drift of the photodiode preamplifier and signal processor by means of a feedback loop in which the threshold voltage follows the envelope of the playback signal, thereby minimizing timing errors within permissible limits for the detector signal. In a dynamic decoupled threshold detector there is no need for threshold voltage adjustment for different discs and players.

Input signal a applied to Summing Amplifier 10 is the analog optical disc playback signal from a dc coupled preamplifier (not shown). Also applied to Summing Amplifier 10 is input b which is a feedback signal that compensates for any thermal drift of the photodiode preamplifier and the signal processor due to temperature or aging. The summed signals a and b from Summing Amplifier 10 are fed to Linear Phase Amplitude Equalizer 11. The signal output from Linear Phase Amplitude equalizer compensates for high frequency response loss in the preceeding stages. This signal is applied to Low Pass Filter 12 in order to filter out high frequency noise. The signal is fed to . Amplifier Driver 13 where it is amplified and provides an output signal c. Signal c is fed to the positive terminal of Comparator 14 and to Peak-to-Peak Detector 15. Peak-to-Peak Detector 15 tracks the positive and negative peaks of signal c and provides two signal outputs f and g. Signal f is the positive peak signal and signal g is the negative peak signal. The signals f and g are fed to Adder Divider 16 where they are added together and then divided by a factor of 2. The output signal h of Adder Divider 16 is the average of signals f and g and it is applied to Low Pass Filter 17 to be filtered. The output signal d from Low Pass Filter 17 is fed to the negative terminal of Comparator 14. Signal d is the tracking threshold signal for accurate signal detection. Comparator 14 compares signal c at its positive terminal with signal d at its negative terminal and generates a thermally compensated digital signal processor output signal e with minimum bias distortion. Signal d from Low Pass Filter 17 is also fed to the positive terminal of feedback Difference Amplifier 18. A reference voltage V" is applied to the negative terminal of Difference Amplifier 18 generating a difference signal which is then fed to Low Pass Filter 19 and then to Summing Amplifier 10 to complete the feedback channel. Low Pass Filter 19 operates at a lower frequency than Low Pass Filter 17. The slowly varying signal b is subtracted from signal a in Summing Amplifier 10.

The dc coupled signal processor of Fig. 7 operates to detect the average of the peak-to-peak analog signal c from Amplifier Driver 13 by means of Peak-to-Peak detector 15 and to pass this average peak-to-peak signal to Low Pass Filter 17 which operates at a cutoff above the periodic drift-like signal. The output of Low Pass Filter 17 is used as a reference voltage for Comparator 14 which is used as a slicer. At the same time, the output of Low Pass Filter 17 is fed to Differential Amplifier 18 with a Low Pass Filter 17 having a cutoff below the frequency of the periodic drift-like signal. This provides signal b which is negatively fed back to Summing Amplifier 10 to compensate for thermal drift. The waveforms of signals a, c, d and e are shown in Figure 11.

The choice of bandwidth of Low Pass Filter 17 needs some discussion. It should have a bandwidth larger than the disc rotational speed so that amplitude modulation due to disc reflectivity variation can be followed by tracking threshold signal d. However, Low Pass Filter 17 should have a low enough bandwidth such that signal d is not sensitive to dropouts of signal c caused by disc defects. Again, the bandwidth of Low Pass Filter 17 should be large so that the time delay of Low Pass Filter 17 is small for signal d to be a better threshold voltage when no dropout occurs. The optimal choice of bandwidth for Low Pass Filter 17 can only be determined by actual experiment. From experiments, a bandwidth of 400 Hz is a good choice for a 4RPS recorder. The threshold voltage, d, is maintained very close to V_r, the reference voltage, by means of the feedback path consisting of the Difference Amplifier 18, Low Pass Filter 19 and Summing Amplifier 10. Thus, all the thermal dc drift of the preamplifier and processor is automatically compensated.

Referring to Fig. 9, there is shown a signal processor essentially the same as Fig. 1 but for means added in order to playback a pregrooved disc. The blocks which are similar to the blocks in Fig. 1 retain the same reference numeral and perform as already described. tn the case of playing back a pregrooved disc, the Sector Gap Detector 30 has applied to its input signal a and it generates a pulse indicating the beginning of the sector header. The output pulse from Sector Gap Detector 30 is fed to the Timing Circuit 31 which generates the correct timing pulse Y as shown in waveform B of Fig. 10. The timing pulse from timing Circuit 31 switches Low Pass Filter 17 to a higher bandwidth such that the threshold signal d can respond quickly to the signal c amplitude change at the junction between the header and data field. Signal Y should end shortly after the beginning of the data field of waveform A.

The present invention provides the advantage in that it accurately converts the readback analog signal of an optical disc from the photodiode preamplifier into a digital signal for different modulation codes with or without dc content. Further, it compensates for thermal dc drift of the photodiode preamplifier, other amplifiers, equalizer and filter. In addition, the playback signal from an optical disc is a band limited signal, having finite rise time and fall time. To convert this analog signal to a digital signal with sharp edges that carry the logical information, a stripping level at the midpoint of the signal is used. Any signal above the stripping level is converted to a logic 1 and below the stripping level to a logic zero. In the circuit, the stripping level dynamically varies with the amplitude of the readback signal. In an optical disc, the reflectivity varies across the surface of the disc generating signal amplitude modulation. Tracking the stripping level to the signal amplitude ensures accurate signal detection. Further advantages are that dropouts do not cause any problem because the stripping voltage time constant is long and at the moment a dropout is over, the voltage is charged up quickly through a very short time constant path. In the pregrooved optical disc, the header data which includes data field gap period, but sync, word sync and track address is prerecorded. During playback, the signal amplitude of the header and the data field is not the same as shown in Fig. 10. There is a step in the amplitude from the header to the data field. The circuit shown in Fig. 9 can detect the playback signal from the pregrooved disc accurately.

From the foregoing, a signal processor for digital optical disc players having means to correct an analog signal from an optical disc with finite rise time to a digital signal with accurately positioned edges by preserving the dc content of the original analog signal has been shown.